iommu.c: rework initialization function

[krystian-hebel]

Test for INVALIDATE_ALL completion is now done inside the initialization function, instead of main.c. Function names were changed to better describe what the function does. Comment describing window in anti-DMA security was updated with new finds.

Two bugs were fixed:

• EventLogInt wasn't properly cleared (it is write-1-to-clear bit)
• CmdBufEn was set on transition to the kernel, which didn't clear it before setting up head/tail pointers, which in turn lead to hang (it is listed as undefined behavior in IOMMU specification)

[krystian-hebel]

SNAFU. Binary built locally from this code (on old Ubuntu, will check newer Fedora) works, while binary built by CI reports wrong self-measurement hash. Other hash values seem to be valid (PCR replay gives current value after fixing just the first entry), so the problem most likely lies in _end_of_measured. I think this might be caused by -R '.note.*'.

EDIT: shouldn't be, _end_of_measured has the same address as base of .note.gnu.property:

[khebel@3M05 secure-kernel-loader]$ objdump skl -h

skl:     file format elf64-x86-64

Sections:
Idx Name          Size      VMA               LMA               File off  Algn
  0 .text         000029a0  0000000000000000  0000000000000000  00001000  2**0
                  CONTENTS, ALLOC, LOAD, READONLY, CODE
  1 .rodata       00000270  00000000000029a0  00000000000029a0  000039a0  2**5
                  CONTENTS, ALLOC, LOAD, READONLY, DATA
  2 .note.gnu.property 00000030  0000000000002c10  0000000000002c10  00003c10  2**3
                  CONTENTS, ALLOC, LOAD, READONLY, DATA
  3 .bss          0000a4d4  0000000000003000  0000000000003000  00004000  2**12
                  ALLOC
  4 .comment      0000002e  0000000000000000  0000000000000000  00003c40  2**0
                  CONTENTS, READONLY
  ( ... more debug sections ... )
[khebel@3M05 secure-kernel-loader]$ objdump skl -t | grep _end_of_measured
0000000000002c10 g       .rodata	0000000000000000 .hidden _end_of_measured

[krystian-hebel]

Local build on Fedora also doesn't give proper hash...

[krystian-hebel]

Something changed in gcc-11, it fails to produce valid SHA1 measurements: #19